Speaker
Description
Charcoal from sedimentary archives is widely used to reconstruct landscape vegetation-fire relationships from modern and ancient landscapes. Digital image analysis is speeding up measurements of sedimentary macro-charcoal, (i.e., grains or particles > ~125—10 000µm). Fire reconstructions now boast higher spatial, temporal, and stratigraphic resolutions. Also, the scope of charcoal analysis has widened to include the determination of fuel-related metrics. However, the reliability and consistency of charcoal metrics across different laboratory processing methods remains untested. There is also underlying variation in how researchers quantify charcoal. Neglecting the study of these key variables may compromise the robustness of charcoal interpretations. Therefore, this initiative funded by the Past Global Changes Inter-Africa mobility grant (PAGES-IAM) aimed to achieve a re-analysis of fire-fuel-type-biomass relationships from charcoal metrics from grassy ecosystems across grazing density, climate, and temporal gradients. We conducted supervised and unsupervised charcoal edge-detection analysis of matched archival imaged samples using Chartool v.1 to from three laboratory methods. We plan to present our preliminary results and discuss how they relate to previous multi-proxy interpretations. We hope to extend this study by developing more training datasets to assist bulk machine-based object-based image analysis (OBIA) to quickly fingerprint the diversity of local and landscape fire regimes from grassy ecosystems
